AI-powered smart cooking assistant that uses computer vision, generative AI, and ingredient-based retrieval to recommend personalized recipes from food users already have at home.
PantryAI is an intelligent cooking assistant that transforms ingredients into personalized meal recommendations using computer vision and generative AI.
Instead of manually searching recipes, users simply upload or scan available ingredients. PantryAI detects the ingredients, maintains an inventory, and generates recipes tailored to the user's pantry.
The project combines modern AI techniques, including image understanding, ingredient recognition, and large language models, to create an intuitive cooking experience.
- Image-based ingredient recognition
- Pantry inventory management
- AI-generated recipe recommendations
- Personalized cooking suggestions
- Ingredient availability matching
- Recipe generation using Google Gemini
- REST API backend
- Modern responsive frontend
- Modular machine learning pipeline
- Upload or scan ingredients
- Computer vision detects available food items
- Pantry inventory is updated automatically
- User requests recipe recommendations
- Gemini generates customized recipes using available ingredients
- Recipes are displayed through the web interface
Image Upload
│
▼
Computer Vision Detection
│
▼
Ingredient Recognition
│
▼
Pantry Inventory
│
▼
Gemini Recipe Generation
│
▼
Personalized Recipe Output
- Python
- JavaScript
- HTML
- CSS
- Google Gemini API
- Computer Vision
- Image Classification
- Ingredient Detection
- Flask
- REST APIs
- JSON
- HTML
- CSS
- JavaScript
- Git
- GitHub
- OpenCV
- VS Code
PantryAI/
│
├── backend/ # Backend API
├── ML/ # Machine learning models
├── pantryAI/ # Frontend application
├── .gitignore
└── package-lock.json
Images are processed using computer vision techniques to identify visible food items.
Detected ingredients are automatically stored in a virtual pantry, allowing users to maintain an up-to-date inventory.
Available ingredients are supplied to Google's Gemini model, which generates complete recipes while prioritizing ingredients already present in the pantry.
Recipes are returned through a clean web interface that minimizes food waste and simplifies meal planning.
Input
Eggs
Milk
Spinach
Cheddar Cheese
Bread
Tomatoes
↓
Generated Recipe
Spinach & Cheddar Breakfast Sandwich
• Toast bread
• Scramble eggs with spinach
• Melt cheddar
• Add sliced tomatoes
• Assemble sandwich
- Modular AI architecture
- End-to-end full-stack application
- Computer vision integration
- LLM-powered recipe generation
- API-driven backend
- Scalable project organization
- Separation of frontend, backend, and ML components
- Artificial Intelligence
- Computer Vision
- Prompt Engineering
- REST API Development
- Full-Stack Development
- Machine Learning Integration
- Python Development
- Software Architecture
- Barcode scanning
- Nutritional analysis
- Grocery list generation
- Meal planning calendar
- User authentication
- Cloud deployment
- Recipe rating system
- Mobile application
- Voice assistant integration
Clone the repository
git clone https://github.com/shri30a/PantryAI.gitMove into the project
cd PantryAIInstall dependencies
pip install -r requirements.txtor
npm installdepending on the component being executed.
Run the backend
python app.pyOpen the frontend and begin using PantryAI.
The objective of PantryAI is to demonstrate how modern AI systems can combine computer vision with large language models to solve a practical consumer problem. By leveraging ingredient recognition and intelligent recipe generation, PantryAI simplifies meal planning while encouraging users to make better use of the food they already own.
This repository is intended for educational and portfolio purposes.
Shrihan Anikapati
Electrical & Computer Engineering (Honors) + Mathematics
The University of Texas at Austin
Interested in AI, Machine Learning, Computer Vision, and Full-Stack Software Engineering.